基于小波变换和FASTICA算法的有噪混合图像盲分离

The Separation of Noisy Mixed Image Based on Wavelet Transform and Independent Component Analysis

简要介绍了独立分量分析的基本数学模型和算法,在此基础上,探讨了独立分量分析在有噪混合图像分离中的应用,提出了一种将小波阈值法去噪与独立分量分析相结合的多通道含噪盲信号分离算法,该算法在对混合含噪图像进行独立分量分析之前,使用小波阈值去噪去除含噪混合图像中的噪声。实验结果表明,该方法能有效地降低噪声信号的影响,较好地恢复了原始图像,解决了传统的独立分量分析方法无法实现加性噪声的多通道含噪盲信号分离的缺陷。

The basic mathematic model and algorithm of independent component analysis was introduced, the application of independent component analysis in separation of noisy mixed images was discussed, and a multiple channel noisy blind signal separation algorithm was proposed, which combines wavelet threshold de-noising and independent component analysis. The algorithm denoises noisy mixed images by wavelet threshold method before independent component analysis. Experimental results show that this method can effectively reduce the effect of noise signal, renew original image, and overcome the limitation that traditional independent component analysis can not solve the additive noisy blind source separation problem.